Proc. Nat!. Acad. Sci. USAVol. 85, pp. 141-145, January 1988Cell Biology

Expression of the insulin-like growth factor II gene in the choroidplexus and the leptomeninges of the adult rat centralnervous system

(in situ hybridization/fetal somatomedin)

FOTINI STYLIANOPOULOU*t, JOSEPH HERBERTt, MARCELO BENTO SOARES*, AND ARGIRIs EFSTRATIADIS*§Departments of *Genetics and Development and tNeurology, Columbia University, 701 West 168th Street, New York, NY 10032

Communicated by Eric R. Kandel, September 11, 1987

ABSTRACT The rat insulin-like growth factor II gene,encoding a fetal somatomedin, expresses a multitranscriptfamily in embryonic/fetal tissues and in the adult brain andspinal cord. By performing in situ hybridization on tissuesections of adult brain and spinal cord, we have found thatthese transcripts are not expressed in neural or glial cells butare expressed in the epithelium of the choroid plexus of eachcerebral ventricle and in the leptomeninges. We propose thatthe choroidal epithelial cells synthesize and secrete insulin-likegrowth factor II into the cerebrospinal fluid.

The somatomedins or insulin-like growth factors I and II(IGF-I and IGF-II) are mitogenic polypeptides that arestructurally similar to proinsulin (for recent reviews, see refs.1-3). We have previously shown that the single-copy rat geneencoding IGF-II (rIGF-II gene) is a complex transcriptionunit that generates a multitranscript family (4). Some of thesetranscripts have been characterized in the rat hepatic BRL-3A cell line that expresses the gene (refs. 4 and 5; M.B.S. andA.E., unpublished results). The gene uses two promoters, P1and P2 (4-6). Transcription from P1 generates the 5' noncod-ing exon E1(1) (1125 nucleotides; nt), whereas transcriptionfrom P2 generates an alternate 5' noncoding exon, E1(2) (94nt). The two predominant mature mRNAs [4.5 and 3.5kilobases (kb)] consist of either E1(1) or E1(2), respectively,connected to three additional exons (E2, 163 nt; E3, 149 nt;and E4, 3.1 kb). Among the several other transcripts, aP2-specific mRNA (1.1 kb) includes only a portion of E4 (thefirst 650 nt) and is generated by differential polyadenylyla-tion.The rIGF-II gene transcripts are present in all fetal or

neonatal tissues that we have examined, but they are ex-tremely rare or undetectable in adult tissues, with theexception of the brain and the spinal cord (4). Similarobservations have been made by other investigators (5, 7-9).To precisely localize rIGF-II gene transcripts in the central

nervous system, we performed in situ hybridization onsections from adult rat brain and spinal cord. We show thatthe IGF-II gene is not expressed in neural or glial cells but isexpressed in the choroid plexus of the cerebral ventricles andthe leptomeninges of the brain and the spinal cord.

MATERIALS AND METHODS

Enzymes and DNAs. Restriction enzymes, Klenow frag-ment of the Escherichia coli DNA polymerase, and M13universal sequencing primers were from New England Bio-labs; SP6 and T7 RNA polymerases and S1 nuclease werefrom Bethesda Research Laboratories; placental RNase

inhibitor (RNasin) and the plasmid vector pGEM-1 (a deriv-ative of plasmid pSP64, containing the promoters of thebacteriophage SP6 and T7 polymerases) were from PromegaBiotec (Madison, WI); [a-32P]dNTPs (800 Ci/mmol; 1 Ci = 37GBq), (a-[35S]thio)UTP (1200 Ci/mmol), and nylon mem-branes (GeneScreenPlus) were from New England Nuclear;pd(N)6 oligomers were from P-L Biochemicals; NTB2 nu-clear track emulsion, D19 developer, and fixer were fromKodak; hybridization probes were prepared from subclonedfragments of the genomic clones rIGF25 and rIGF15, and thecDNA clone 27 (see ref. 4).RNA Analysis. Total cell RNA was prepared from BRL-3A

cells (American Type Culture Collection Cell RepositoryLine 1442) and dissected choroid plexi or other brain regionsby the guanidine thiocyanate/cesium chloride procedure(10). For RNA blot hybridization, RNA was electrophoresedon formaldehyde/1% agarose gels and then transferred ontonylon membranes. Prehybridization and hybridization wereas described (11). The hybridization probes were eitheruniformly 32P-labeled, mRNA-complementary, single-stranded DNA synthesized on M13 templates (12) or DNAfragments labeled by randomly primed synthesis usingpd(N)6 primers, [a-32P]dNTPs, and Klenow enzyme as de-scribed (13). S1 nuclease-protection analysis was done asdescribed (4).In Situ Hybridization. Tissues from Sprague-Dawley rats

were sectioned and prepared for in situ hybridization asdescribed (14). Hybridization, followed by washing understringent conditions, autoradiography, exposure to photo-emulsion, development, and counterstaining with hematox-ylin/eosin were as described (15). The RNA hybridizationprobes were prepared as follows. A 545-base pair (bp) codingregion fragment from rIGF-II cDNA clone 27 (4) was isolat-ed. This fragment extends between aBamHI site located onenucleotide downstream from the ATG initiator and an EcoRIlinker present at the end of the fragment three nucleotidesdownstream from the TGA terminator. The fragment wassubcloned into the BamHI and EcoRI sites of the pGEM-1vector polylinker. Transcription with 17 polymerase fromBamHI-linearized plasmid generated antisense probe,whereas transcription with SP6 polymerase from EcoRI-linearized plasmid generated sense (control) probe. Tran-scription reactions with these polymerases (10-,A vol each)were performed according to the manufacturer's specifica-tions, using 25 ,uM 35S-UTP, 500 ttM each of the other threedNTPs, 20 units of RNasin, 1 gug of linearized template, and20 units of polymerase.

Abbreviations: IGF-I and -II, insulin-like growth factors I and II,respectively; rIGF-II, rat IGF-II; CSF, cerebrospinal fluid.tPermanent address: School of Health Sciences, University ofAthens, Greece.§To whom reprint requests should be addressed.

141

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

4, 2

021

142 Cell Biology: Stylianopoulou et al.

RESULTSWe had previously demonstrated by RNA blot hybridizationthat rIGF-II gene transcripts are present in the steady-statepopulation ofRNA extracted from embryonic/fetal and adultbrain and from adult spinal cord (4). To localize these RNAspecies in the nervous system of adult rats we performed insitu hybridization studies. Sagittal and coronal sections ofwhole rat brain at 50-Am intervals were hybridized to 35S_labeled, single-stranded, coding-region RNA probes (senseand antisense; see Materials andMethods). Autoradiographyof the sections that were hybridized with antisense proberevealed intense hybridization signals localized within eachof the four cerebral ventricles, whereas there was no evi-dence of hybridization above background in the remainder ofthe brain tissue (Fig. 1 a-d). Hybridization signal outliningthe edges of the sections was also detected (Fig. 1 b-d). Thecontrol (sense) RNA probe yielded negative results (Fig. 1 c'and d'). These observations suggested that the rIGF-II geneis active in the choroid plexus of the ventricles and in theleptomeninges.The choroid plexus is a villous structure that projects into

each cerebral ventricle (refs. 16-21 and refs. therein) and islargely responsible for the production and secretion of thecerebrospinal fluid (CSF). The plexus consists of choroidalepithelium (a single layer ofcuboidal cells) covering a stromalcore containing blood vessels (Fig. le). The epithelium ismodified ependyma (the lining of the ventricular cavities),whereas the core is thought to be derived from pia mater.The leptomeninges (Fig. le) are traditionally described as

two membranes, the arachnoid mater and the pia mater,separated by the subarachnoid space, which is filled withCSF. They are composed predominantly of flat leptomenin-geal cells, organized in different locations into cell layers ofvariable thickness (refs. 16, 22-28, and refs. therein). Gen-erally, the arachnoid is several (usually three) cell layersthick, whereas the pia consists of a single layer of cells. Somecells of the inner layer of the arachnoid are connected to cellsof the pia by thin cytoplasmic processes (trabeculae), whichtraverse the subarachnoid space (Fig. le). Pial cells accom-pany the walls of the larger blood vessels as they enter thecerebral parenchyma from the subarachnoid space, therebyforming a perivascular coat (Fig. le).

Histological examination of the brain sections after expo-sure to a nuclear track emulsion confirmed our interpretationof the autoradiographic results: numerous silver grains werepresent throughout the cytoplasm of the choroidal epithelialcells (Fig. 2 a-c) at approximately the same density in theplexus of each ventricle. However, we cannot exclude fromthese data the contribution of stromal cells to the hybridiza-tion signal. In contrast, silver grains above background wereabsent from the ependymal cells; a sharp transition betweenhybridizing and nonhybridizing epithelium was evident at thepoint of attachment of the choroid plexus to the ependymallining (Fig. 2 a and b).

Uniformly distributed silver grains at high density werealso detected in the leptomeninges covering the cerebralhemispheres (Fig. 2 d-f), the cerebellum, and the brain stem(data not shown). Notably, the cytoplasm of the trabeculae,which connect the arachnoid and pial membranes, alsoshowed intense hybridization (Fig. 2f). Although we couldnot distinguish histologically between leptomeningeal cellsand arachnoidal fibroblasts, the prominent trabecular hybrid-ization strongly suggests that at least the former cells containIGF-II transcripts. We found that assessment of silver grainsover the pia mater was more difficult than for the arachnoidbecause the pia is thinner and can be easily distorted. Thus,in some regions we did not detect hybridization abovebackground along the parenchymal perimeter correspondingto the pia mater, but in other regions silver grains were clearly

a

b .

C -'.--

E. .

d . -

e

C

di

FIG. 1. Autoradiograms of brain sections hybridized with "5S-labeled RNA probes, antisense (a-d) or sense (c' and d'). Thehybridization signal is localized within one ofthe lateral ventricles (atleft) and the fourth ventricle (at right) seen in sagittal section (a); andwithin the lateral (left and right) and third (center) ventricles (b-d)seen in coronal sections at the levels of the anterior third ventricle(b), foramen of Monroe (c), and posterior third ventricle (d).Hybridization signal is outlining the perimeter of the sections and theinterhemispheric fissure in b-d. (c' and d') Sections adjacent to c andd, respectively. (b') Diagram of the section in b. The squares A-Findicate, respectively, the regions shown in a-fof Fig. 2. An arrowindicates the transverse cerebral fissure, the perimeter of which isoutlined by hybridization signal in b. (e) Schematic representation ofthe relationships between the meninges, choroid plexus, ependyma,and brain parenchyma.

present, albeit at lower density than in the arachnoid.Additional evidence, suggesting that cells of the pia layer ofthe leptomeninx also synthesize IGF-II transcripts, wasprovided by the observation that many larger-diameterintraparenchymal blood vessels were encircled by a ring ofsilver grains presumably corresponding to the pial perivas-

Proc. NatL Acad. Sci. USA 85 (1988)

4.

tv

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

4, 2

021

Proc. Natl. Acad. Sci. USA 85 (1988) 143

I -lIIEll1gareBi ii Tr .:.;

'4ii

FIG. 2. In situ hybridization of regions of brain sections (see Fig. lb'). Silver grains are seen in the epithelial cells of the choroid plexus (CP)of the lateral ventricle (a-c) and in the leptomeninx (L) or its arachnoid component (A) (d-f). The abrupt loss of hybridization signal in theependymal (EP) cells is indicated by an arrowhead (a and b; different sections of the same region in different magnification). V, P, and T indicatethe ventricular space, the parenchyma, and trabeculae. Arrows in e indicate rings of hybridization signal in the perivascular coat (PVC) ofparenchymal vessels (see also inset e').

cular coat (Fig. 2e and inset e'). Autoradiographic andhistological examination of sections of the spinal cord alsodemonstrated an intense hybridization signal localized ex-clusively in the leptomeninx (data not shown).

Previously, rIGF-II gene transcripts were detected in totalRNA extracted from various brain regions by RNA blothybridization using coding region probes (4, 5, 7). Theseresults should now be interpreted as due to the presence inthe preparations ofRNA from the choroid plexus and/or theleptomeninx, portions of which presumably remained asso-ciated with parenchyma during tissue dissection. This inter-pretation is consistent with estimates of the level of IGF-IItranscripts by S1 nuclease-protection analysis; the concen-tration of IGF-II RNA species is -200-fold higher in totalRNA extracted from dissected choroid plexi than in RNAprepared from parenchyma of various brain regions stillassociated with the leptomeninges (data not shown). Unfor-tunately, microdissection of the leptomeninges away fromneural parenchyma was difficult in our attempt to prepareRNA derived exclusively from neural tissue for comparativeanalysis. Nevertheless, should neural cells also express therIGF-II gene, the level oftranscripts must be below detectionlimits by in situ hybridization, a method with adequate

sensitivity to easily detect 150-300mRNA molecules per cell(29) and probably as few as 5-10 molecules per cell (30).The conclusion that the parenchymal cells do not actively

transcribe the rIGF-II gene is based on our assessment ofsilver-grain density in coronal and sagittal brain sections. Forcomparison we examined five different microscopic fields ofparenchyma from different brain regions in four pairs ofadjacent sections hybridized with antisense or sense probe.In each field, grains were counted in five different areas, eachcorresponding to 3.77 ,um2. The mean ± SD of the calculatedratios of the experimental and control values for each fieldwas 1.02 ± 0.06, which demonstrates that the appearance ofgrains in neural parenchyma is due to background hybrid-ization. The corresponding value in parallel examination ofthe choroid plexus was 19.34 ± 3.55.The mechanism that maintains the activity of the rIGF-II

gene in the adult choroid plexus is not known. However, wehave excluded the operation of developmental stage-specificpromoters (see Discussion). Detailed analysis of the blothybridization profiles of IGF-II transcripts in choroid plexusRNA, using not only coding region probe but also probesfrom the alternate 5' noncoding regions and the 3' noncodingregion, demonstrated transcription from both the P1 and P2promoters of the gene and did not reveal the occurrence of

Cell Biology: Stylianopoulou et al.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

4, 2

021

144 Cell Biology: Stylianopoulou et al.

any transcripts different from those observed previously (4)in BRL-3A cells and in rat fetal tissues (data not shown).

DISCUSSION

IGF-II is clearly an embryonic/fetal somatomedin in the rat(31), although its exact developmental role is still not known.IGF-II transcripts are expressed during the neonatal period,but these transcripts disappear from most adult rat tissues (4,9, 32). Although transcripts from the IGF-II gene have alsobeen detected by in situ hybridization in several human fetaltissues (33), the developmental expression of this geneapparently differs between humans and rats. In humans,IGF-II appears present in serum at lower concentrationsduring fetal development than in adults (34). Moreover,IGF-II transcripts are certainty synthesized in human liverduring adult life as a consequence of the operation of'an adultpromoter (35). We have been unable thus far to identify acorresponding adult promoter in rats, and we do not knowwhether it exists. However, our analysis indicates that theappearance of rIGF-II transcripts in the choroid plexus andthe leptomeninges during adult life is due to the operation ofthe same two promoters (P1 and P2) that function during theembryonic stages.During development, the cells of the choroidal epithelium

differentiate distinctly from the ependymal cells, despitecommon embryological origin from the neuroepithelial cellsthat line the internal surface of the neural tube (18). Theirmorphological and functional differences can now be corre-lated with the presence or absence of two molecular markers;choroidal, but not ependymal, cells transcribe the rIGF-IIand transthyretin genes. Transthyretin (prealbumin), a trans-port protein for retinol and thyroxin, is synthesized de novoin the choroidal epithelium and then secreted into the CSF(36-39).Our evidence, demonstrating that the rIGF-II gene is active

in the epithelial cells of the choroid plexus and in thepia/arachnoid cells throughout the neuraxis, establishes acommon biochemical marker for these cell types, althoughtheir embryological relationship is uncertain. The origin ofthe leptomeningeal cells is still a matter of controversy. It isbelieved that they are derived from neuroectoderm (possiblythe neural crest), or from mesoderm, or from both (see refs.27, 40-43, and refs. therein). Nevertheless, the plexus andthe leptomeninges are linked physiologically as a functionalunit. The CSF is produced and secreted into the ventricles bythe choroid plexus, exits through foramina in the roof of thefourth ventricle, percolates through the subarachnoid space,and is reabsorbed into the venous system via the arachnoidvilli (see Fig. le). Thus, the plexus and the leptomeningesconstitute the major interfaces of the blood and CSF com-partments (16).The presence of IGF-II in human CSF has been demon-

strated (44), and its CSF-to-plasma ratio has been shown tobe significantly higher than that expected if IGF-II weretransported into CSF via nonspecific routes (44, 45). Inhumans the IGF-II gene is also probably expressed in thechoroid plexus. Interestingly, the IGFs, which are found inthe circulation associated with binding proteins (46), are alsobound to carrier polypeptides in the CSF (47). One of thesecarriers, a 34-kDa polypeptide, is enriched in the CSF ascompared with plasma and exhibits selective affinity forIGF-II (47). From these observations in conjunction with ourresults, we propose that the choroid plexus synthesizes andsecretes IGF-II into the CSF.

IGF-II has been detected by radioimmunoassays or radio-receptor assays in adult human brain regions (48, 49). Theseresults (48) indicated that the concentration of IGF-II was onthe average at least 10 times higher in the pituitary than in anybrain region. In the rat we observed abundant IGF-II tran-

scripts in the embryonic pituitary by in situ hybridization(unpublished results), but we were unable to detect any suchtranscripts in the pituitary of adult animals by RNA blothybridization or in situ hybridization (data not shown).The detection oftranscriptional activity ofthe rIGF-II gene

in both the arachnoid and pial cells, which cannot bediscriminated morphologically (22-25), suggests that theyshare at least some common function. However, our datacannot demonstrate whether the IGF-II gene is active in allof the leptomeningeal cells, or in a specific subset [at leasttwo, and possibly more, cellular types constitute the cellpopulation of the leptomeninges (23, 24, 42)]. Also, ourresults do not prove that these cells are secretory, althoughthis function is not unlikely. Comparative examination ofarachnoid and meningioma cells has suggested, from a seriesof morphological and biochemical criteria, that such cellsexhibit both epithelial and mesenchymal features (42). Thesecretory nature of tumor cells in some meningiomas, whichare of leptomeningeal origin, has been suggested (50).The local production of IGF-II in structures associated

with the embryonic and adult nervous system and its circu-lation in the CSF suggest that this growth factor is function-ally important for neurons and/or glial cells. Receptors forboth IGF-I and IGF-II are present in adult human and ratbrain (51). However, the function of IGF-II in the centralnervous system is not known; and direct immunocytochem-ical evidence for the localization of the growth factor innervous tissue is currently not available. Nevertheless, theidentification of IGF-II receptors in the brain implies that thechoroid plexus, and possibly the leptomeninges, may con-stitute a paracrine system. Several observations are consis-tent with this notion. For example, a glia maturation factor(an acidic 14-kDa polypeptide), which has no mitogenic effectin the absence of serum, can stimulate the growth rate ofnormal astroblasts cultured in a defined serum-free mediumcontaining physiological concentrations of IGF-II (52). More-over, IGF-II stimulates the incorporation of tritiated thymi-dine into DNA in primary cultures of fetal rat brain cells (53)and in a cell line of human neuroblastoma cells that prolif-erate in the presence of the mitogen in a defined medium (54).Finally, IGF-II enhances neurite outgrowth in the samecloned human neuroblastoma cells (55) and in primarycultures of sensory and sympathetic neurons from embryonicchick ganglia (56, 57).

We thank Luis Cucuta for expert technical assistance and EricSchon, John Pintar, and Debra Wolgemuth for use of facilities. Thiswork was supported by a grant from the National Institutes of Healthto A.E., a gift to the laboratory of A.E. from the Bristol-Myers Co.,a Clinical Investigator Development Award from the NationalInstitute of Neurological and Communicable Diseases and Stroke toJ.H., and awards from the Dana and Diamond Foundations to J.H.

1. Froesch, E. R., Schmid, C., Schwander, J. & Zapf, J. (1985)Annu. Rev. Physiol. 47, 443-467.

2. Zapf, J. & Froesch, E. R. (1986) Horm. Res. 24, 121-130.3. Baxter, R. C. (1986) Adv. Clin. Chem. 25, 49-115.4. Soares, M. B., Turken, A., Ishii, D., Mills, L., Episkopou, V.,

Cotter, S., Zeitlin, S. & Efstratiadis, A. (1986) J. Mol. Biol.192, 737-752.

5. Frunzio, R., Chiariotti, L., Brown, A. L., Graham, D. E.,Rechler, M. M. & Bruni, C. B. (1986) J. Biol. Chem. 261,17138-17149.

6. Evans, T., DeChiara, T. & Efstratiadis, A. (1987) J. Mol. Biol.,in press.

7. Lund, P. K., Moats-Staats, B., Hynes, M. A., Simmons,J. G., Jensen, M., D'Ercole, A. J. & Van Wyk, J. J. (1986) J.Biol. Chem. 261, 14539-14544.

8. Graham, D. E., Rechler, M. M., Brown, A. L., Frunzio, R.,Romanus, J. A., Bruni, C. B., Whitfield, H. J., Nissley, S. P.,Seelig, S. & Berry, S. (1986) Proc. NatI. Acad. Sci. USA 83,4519-4523.

Proc. Natl. Acad. Sci. USA 85 (1988)

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

4, 2

021

Proc. Natl. Acad. Sci. USA 85 (1988) 145

9. Brown, A. L., Graham, D. E., Nissley, S. P., Hill, D. J.,Strain, A. J. & Rechler, M. M. (1986) J. Biol. Chem. 261,13144-13150.

10. Chirgwin, J. M., Przybyla, A. E., MacDonald, R. J. & Rutter,W. J. (1979) Biochemistry 18, 5294-5299.

11. Zeitlin, S. & Efstratiadis, A. (1984) Cell 39, 589-602.12. Hu, N. & Messing, J. (1982) Gene 17, 271-277.13. Feinberg, A. P. & Vogelstein, B. (1983) Anal. Biochem. 132,

6-13.14. Gee, C. E. & Roberts, J. L. (1983) DNA 2, 157-163.15. Fremeau, R. T., Lundblad, J. R., Pritchett, D. B., Wilcox,

J. N. & Roberts, J. L. (1986) Science 234, 1265-1269.16. Davson, H. (1967) Physiology of the Cerebrospinal Fluid

(Churchill Livingstone, Edinburgh).17. Tennyson, V. M. & Pappas, G. D. (1968) Prog. Brain Res. 29,

63-85.18. Netsky, M. G. & Shuangshoti, S. (1975) The Choroid Plexus in

Health and Disease (Univ. Press of Virginia, Charlottesville,VA).

19. Rodriguez, E. M. (1976) J. Endocrinol. 71, 406-443.20. Agnew, W. F., Alvarez, R. B., Yuen, T. G. H. & Crews,

A. K. (1980) Cell Tissue Res. 208, 261-281.21. Keep, R. F., Jones, H. C. & Cawkwell, R. D. (1986) Dev.

Brain Res. 27, 77-85.22. Pease, D. C. & Schultz, R. L. (1958) Am. J. Anat. 102,

301-321.23. Waggener, J. D. & Beggs, J. (1967) J. Neuropathol. Exp.

Neurol. 26, 412-426.24. Morse, D. E. & Low, F. N. (1972) Am. J. Anat. 133, 349-368.25. Dermietzel, R. (1975) Cell Tissue Res. 164, 309-329.26. McLone, D. G. & Bondareff, W. (1975) Am. J. Anat. 142,

273-294.27. Schachenmayr, W. & Friede, R. L. (1978) Am. J. Pathol. 92,

53-68.28. Krisch, B., Leonhardt, H. & Oksche, A. (1983) Cell Tissue

Res. 228, 597-640.29. Wuenschell, C. W., Fisher, R. S., Kaufman, D. L. & Tobin,

A. J. (1986) Proc. Natl. Acad. Sci. USA 83, 6193-6197.30. Siegel, R. E. & Young, W. S. (1986) in In Situ Hybridization in

the Brain, ed. Uhl, G. (Plenum, New York), pp. 63-72.31. Moses, A. C., Nissley, S. P., Short, P. A., Rechler, M. M.,

White, R. M., Knight, A. B. & Higa, 0. Z. (1980) Proc. Natl.Acad. Sci. USA 77, 3649-3653.

32. Murphy, L. J., Bell, G. I. & Friesen, H. G. (1987) Endocrinol-ogy 120, 1279-1282.

33. Han, V. K. M., D'Ercole, A. J. & Lund, P. K. (1987) Science236, 193-197.

34. Ashton, I. K., Zapf, J., Einschenk, I. & MacKenzie, I. Z.(1985) Acta Endocrinol. 110, 558-563.

35. de Pagter-Holthuizen, P., Jansen, M., van Schaik, F. M. A.,van der Kammen, R., Oosterwijk, C., Van den Brande, J. L. &Sussenbach, J. S. (1987) FEBS Lett. 214, 259-264.

36. Kato, M., Soprano, D. R., Makover, A., Kato, K., Herbert, J.& Goodman, D. (1986) Differentiation 31, 228-235.

37. Herbert, J., Wilcox, J. N., Pham, K.-T. C., Fremeau, R. T.,Zeviani, M., Dwork, A., Soprano, D. R., Makover, A., Good-man, D., Zimmerman, E. A., Roberts, J. L. & Schon, E. A.(1986) Neurology 36, 900-911.

38. Dickson, P. W., Aldred, A. R., Marley, P. D., Guo-Fen, T.,Howlett, J. & Schreiber, G. (1985) Biochem. Biophys. Res.Commun. 127, 890-895.

39. Dickson, P. W., Aldred, A. R., Marley, P. D., Bannister, D.& Schreiber, G. (1986) J. Biol. Chem. 261, 3475-3478.

40. Gil, D. R. & Ratto, G. D. (1973) Acta Anat. 85, 620-623.41. Morse, D. E. & Cova, J. L. (1984) Anat. Rec. 210, 125-132.42. Schwechheimer, K., Kartenbeck, J., Moll, R. & Franke,

W. W. (1984) Lab. Invest. 51, 584-591.43. Rutka, J. T., Giblin, J., Dougherty, D. V., McCulloch, J. R.,

DeArmond, S. J. & Rosenblum, M. L. (1986) J. Neuropathol.Exp. Neurol. 45, 285-303.

44. Haselbacher, G. & Humbel, R. (1982) Endocrinology 110,1822-1824.

45. Pardridge, W. M. (1986) Ann. N.Y. Acad. Sci. 481, 231-249.46. Binoux, M., Hossenlopp, P., Hardouin, S., Seurin, D., Las-

sarre, D. & Gourmelen, M. (1986) Horm. Res. 24, 141-151.47. Hossenlopp, P., Seurin, D., Segovia-Quinson, B. & Binoux,

M. (1986) FEBS Lett. 208, 439-444.48. Haselbacher, G. K., Schwab, M. E., Pasi, A. & Humbel,

R. E. (1985) Proc. Natl. Acad. Sci. USA 82, 2153-2157.49. Carlsson-Skwirut, C., Jornvall, H., Holmgren, A., Andersson,

C., Bergman, T., Lundquist, G., Sjogren, B. & Sara, V. R.(1986) FEBS Lett. 201, 46-50.

50. Alguacil-Garcia, A., Pettigrew, N. M. & Sima, A. A. F. (1986)Am. J. Surg. Pathol. 10, 102-111.

51. Gammeltoft, S., Haselbacher, G. K., Humbel, R. E.,Fehlmann, M. & Van Obberghen, E. (1985) EMBO J. 4,3407-3412.

52. Lim, R., Miller, J. F., Hicklin, D. J., Holm, A. C. & Ginsberg,B. H. (1985) Exp. Cell Res. 159, 335-343.

53. Enberg, G., Tham, A. & Sara, V. R. (1985) Acta Physiol.Scand. 125, 305-308.

54. Mattsson, M. E. K., Enberg, G., Ruusala, A.-I., Hall, K. &Pahlman, S. (1986) J. Cell Biol. 102, 1949-1954.

55. Recio-Pinto, E. & Ishii, D. N. (1984) Brain Res. 302, 323-334.56. Bothwell, M. (1982) J. Neurosci. Res. 8, 225-231.57. Recio-Pinto, E., Rechler, M. M. & Ishii, D. (1986) J.

Neurosci. 6, 1211-1219.

Cell Biology: Stylianopoulou et al.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

4, 2

021